基于RBF神经网络和模型在线更新的风速估计方法  

WIND SPEED ESTIMATION METHOD BASED ON RBF NEURAL NETWORKS AND MODEL ONLINE UPDATE

作　　者：曹忠鹏 陈文婷[1,2] 艾超[1] 王秦未[1] 张晨阳 杜泽莉 Cao Zhongpeng;Chen Wenting;Ai Chao;Wang Qinwei;Zhang Chenyang;Du Zei(State Key Laboratory of Crane Technology,Yanshan University,Qinhuangdao 066004,China;State Key Laboratory of Fluid Power and Mechatronic Systems,Zhejiang University,Hangzhou 310030,China)

机构地区：[1]燕山大学起重机械关键技术全国重点实验室,秦皇岛066004 [2]浙江大学流体动力基础件与机电系统全国重点实验室,杭州310030

出　　处：《太阳能学报》2024年第10期453-458,共6页Acta Energiae Solaris Sinica

基　　金：国家自然科学基金青年项目(52205071);中央引导地方科技发展专项资金(ZYYD2022C09);流体动力与机电系统国家重点实验室开放基金课题(GZKF-202110);先进制造成形技术及装备国家地方联合工程研究中心育苗项目。

摘　　要：基于NREL5MW风力发电机原模型和简化模型,进行流固耦合和计算流体力学仿真,研究结果表明风力发电机桨距角和叶尖速比影响风轮上游来流风速演化。结合来流风速演化和推力系数理论模型表明,通过推力更新诱导系数,进而在线更新风速演化模型,能提高风速估计精度。考虑推力时变性和非定常性,提出以桨距角和叶尖速比为输入,基于RBF神经网络的推力逼近方法,进而提出基于模型在线更新的风轮平面有效风速估计方法。最后通过激光雷达现场测风数据进行验证,结果表明,基于RBF神经网络和模型在线更新的风速估计方法较基于时间信息提取算法和来流风速演化模型的风速估计方法精度提高15.5%。The influence of the pitch angle and tip speed ratio on wind speed evolution in the induction zone of the wind turbine incoming flow is demonstrated through fluid structure interaction and computational fluid dynamics simulation respectively,based on both the original NREL 5 MW wind turbine model and its simplified one.Combined with the theoretical model of wind speed evolution and thrust coefficient,it can be obtained that changing the induction factor in the thrust model and then updating the thrust coefficient in the wind speed evolution model online can improve the accuracy of the effective wind speed estimation.Considering the time-varying and non-stationary nature of thrust,a thrust approximation method based on RBF neural networks with the input of pitch angle and tip speed ratio is put forward.Then,an estimation method of wind turbine plane effective wind speed based on online model updating is proposed.At last,validation is conducted using on-site wind measurement data from LiDAR,and the results show that the wind speed estimation method based on RBF neural networks and model online update improves the estimation accuracy by 15.5%compared to the estimation method with time information extraction algorithm and incoming flow wind speed evolution model.

关 键 词：风力发电机 流固耦合 RBF神经网络 桨距角 叶尖速比 有效风速 

分 类 号：TK89[动力工程及工程热物理—流体机械及工程]